How do i calculate weights for max pooling output?

For example if there are 10 inputs, a pooling filter of size and a stride 2, how many weights including bias are required for the max pooling output ?

  • $\begingroup$ Do you mean pooling filter of size 2 and a stride of 2, or did you forget to add the size number? I'm just making sure I understand the exact question $\endgroup$ – Krrrl Oct 27 '19 at 21:07
  • $\begingroup$ yup added, thanks $\endgroup$ – Tibby Oct 27 '19 at 21:25
  • $\begingroup$ Pooling layers don't have any trainable weights. They just perform an aggregation operation on their input. $\endgroup$ – Djib2011 Oct 27 '19 at 21:36
  • $\begingroup$ Also, now that I read the question again, I am unsure what you are asking. What exactly are you asking/looking for? Pooling layers in CNN have no learnable weights, because their sole function is to decrease the inputs dimensions. Are you asking how to compute the backpropagation(gradient) for the layer that comes before a pooling layer? $\endgroup$ – Krrrl Oct 27 '19 at 21:36
  • $\begingroup$ Got it, i think i confused a pooling layer with a normal filter. Your response clarified things for me. thank you $\endgroup$ – Tibby Oct 27 '19 at 21:38

A Max Pool layer don't have any trainable weights. Only hyperparameters is present and they are non-trainable. The max pooling process calculates the maximum value of the filter, which consists of no weights and biases. It is purely a way to down scale the data to a smaller dimension. Hope this helps you and have a nice day!

| improve this answer | |

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.